Underwriter assess and score risks from insurance contract documents using Azure Open AI

Question

Underwriter assess and score risks from insurance contract documents using Azure Open AI

Janarthanan S 700

I work for insurance client request where to use upload contract documents in the Azure BLOB Storage and extraction is done through Azure Document Intelligence for pre processing tasks. My task is to summarize the contract documents. How to summarize the documents using Azure OpenAI service? in order to ensure the underwriter decision is correct or not?

Regards,

Janarthanan S

dupammi 8,615 Reputation points Microsoft External Staff

2023-09-21T14:25:09.24+00:00

Hi @Janarthanan S ,

Thanks for using Microsoft Q&A platform. I understand that you are trying to upload contract documents in Azure Blob Storage, perform pre-processing with Azure document intelligence, Summarize the documents using Azure OpenAI service and do decision making on the summaries. I will be happy to assist you with this.

Here is the detailed information related to the document summarization, which can be run from Form Recognizer to OCR documents.

This can also generate summaries for extracted documents. It also shows how to generate one short summary for each document.

Please refer below for the implementation details:

https://github.com/Azure/azure-openai-samples/blob/main/fundamentals/langchain/02_large_document-summerization.ipynb

I hope this information helps!

Answer accepted by question author

0 additional answers

Your answer

dupammi 8,615 Reputation points Microsoft External Staff

2023-09-21T14:25:09.24+00:00

Hi @Janarthanan S ,

Thanks for using Microsoft Q&A platform. I understand that you are trying to upload contract documents in Azure Blob Storage, perform pre-processing with Azure document intelligence, Summarize the documents using Azure OpenAI service and do decision making on the summaries. I will be happy to assist you with this.

Here is the detailed information related to the document summarization, which can be run from Form Recognizer to OCR documents.

This can also generate summaries for extracted documents. It also shows how to generate one short summary for each document.

Please refer below for the implementation details:

https://github.com/Azure/azure-openai-samples/blob/main/fundamentals/langchain/02_large_document-summerization.ipynb

I hope this information helps!

Answer 1

Azar 31,705 MVP Volunteer Moderator

Hi @Janarthanan S

Start by uploading your insurance contract documents to Azure Blob, then StorageSet up Azure Document Intelligence to extract key information and preprocess your contract documents.

Retrieve the extracted information from Azure Document Intelligence.

use Azure OpenAI's text summarization capabilities into your workflow. You can leverage models like GPT-3, which can be fine-tuned for specific tasks.

Fine-tune the Azure OpenAI model using a dataset of summarized contract documents, and finally Develop a script or application that interfaces with the Azure OpenAI API. Provide the extracted text from your contract documents as input and request summaries from the fine-tuned model.

If you find this answer usefull kindly upvote and accept for any assistance ping here thanks much.

Janarthanan S 700 Reputation points

2023-09-21T08:22:35.4666667+00:00

Thanks for the reply Azar and can you please provide some sample code snippets.

Regards,

Janarthanan S

Azar 31,705 MVP Volunteer Moderator

sure dont forget to custamise the code before using

import openai

# Configure your API key
api_key = "YOUR_API_KEY_HERE"

# Text to be summarized
input_text = """
Insert your insurance contract document text here. Make sure it's properly extracted from Azure Document Intelligence.
"""

# Use the OpenAI API for text summarization
response = openai.Completion.create(
    engine="text-davinci-003",  # You can also use "gpt-3.5-turbo" for faster results
    prompt=f"Summarize the following insurance contract:\n{input_text}",
    max_tokens=150,  # Adjust the maximum length of the summary as needed
    n=1,  # Number of responses to generate
    stop=None,  # You can specify stop words if needed
    temperature=0.7,  # Adjust the temperature for creativity (0.2 to 1.0)
    api_key=api_key,
)

# Extract the summarized text from the response
summary = response.choices[0].text.strip()

# Print the generated summary
print("Summarized Contract:")
print(summary)

Share via

Underwriter assess and score risks from insurance contract documents using Azure Open AI

0 additional answers

Your answer